The Trump administration’s long-awaited AI action plan is here, and it’s hitting notes likely familiar to watchers of the administration’s AI policy so far: slashes to “onerous regulation,” opposition to “woke AI,” antagonism toward states that regulate AI, and fostering “global AI dominance.”
The administration released the plan this week to follow through on a deadline it set in January, when Trump repealed former President Biden’s AI executive order shortly after his inauguration. The plan encompasses a set of policy guidelines along with three executive orders covering expansion of tech exports to allies, speeding data center construction, and attempting to curb ideological bias in models.
Here are some takeaways:
The tech industry cheered the plan, while safety groups criticized it
Tech executives like Nvidia CEO Jensen Huang and AMD’s Lisa Su reportedly spoke in support of the plan at an unveiling event in Washington, DC, the Wall Street Journal reported.
Tech industry trade group NetChoice, which counts Amazon, Google, and Meta as members, also welcomed the plan.
“NetChoice applauds the White House’s AI Action Plan overall and is encouraged to see the focus on red tape reduction and investment in America’s future,” Patrick Hedger, NetChoice’s director of policy, said in a statement.
Meanwhile, nonprofits focused on AI safety criticized the plan as overly friendly to the tech industry. A coalition consisting of more than 90 advocacy groups and labor unions released its own “People’s AI Action Plan” that sought to counter the administration’s agenda.
“The White House AI Action Plan is written by Big Tech interests invested in advancing AI that’s used on us, not by us,” Sarah Myers West and Amba Kak, co-executive directors at the AI Now Institute, wrote in a statement.
A federal moratorium on state regulation resurrected
A moratorium that would have banned state and local regulation of AI for a decade may have been resoundingly voted down by the Senate in early July, but the goal of limiting state power lives on. The plan threatens to withhold AI-related funding from states that enforce “burdensome AI regulations,” but also acknowledges their right to “pass prudent laws.”
California State Senator Scott Wiener, the author behind an AI safety bill that’s wending through the state’s legislature now, characterized this as “bullying” in an otherwise mixed appraisal.
“I’m happy to see there’s a bipartisan consensus around the need to accelerate innovation and bolster security and safety research into the most powerful AI systems,” Wiener wrote in a statement to Tech Brew. “At the same time, there’s a lot that’s bad in this plan, like bullying states into not regulating AI.”
Open-weights models get a shoutout
The Trump administration weighed in on the ongoing split between open-weight and closed-source models, saying the government should support leading open models. Many of its recommendations are about expanding access to computing power, such as the National AI Research Resource established under Biden.
Workers could be retrained to work at data centers
The plan’s answer to AI-related job losses centered on retraining workers to fill roles like HVAC technicians and electricians, in the coming wave of data centers it plans to fast-track. The plan also called for “a serious workforce response to help workers navigate [the AI] transition.”